The invention relates generally to technology for sensing and recording fingerprints and, more particularly to systems, devices and methods for fingerprint motion tracking alone and in combination with fingerprint image processing.
A number of devices and techniques exist for sensing, capturing, and reconstructing the image of a fingerprint as it moves across a sensor array. Though many devices exist to sense and record an entire fingerprint, partial fingerprint sensing devices have been developed for small portable devices to save space. The sensing devices themselves vary widely, and many devices and related techniques exist for sensitively detecting the presence of the finger surface and features located on the surface that make up the unique fingerprint of a person. For example, one common configuration used for a fingerprint sensing surface includes CCD (charge coupled devices) or C-MOS circuits. These components are embedded in a sensing surface to form a matrix of piezoelectric elements that generate signals in response to pressure applied to the surface by a finger. These signals are read by a processor and used to reconstruct the fingerprint of a user and to verify identification. Other devices include a matrix of optical sensors that read light reflected off of a person's finger and onto optical elements The reflected light is converted to a signal that defines the fingerprint of the finger analyzed and is used to reconstruct the fingerprint and to verify identification. More modern devices include static or radio frequency (RF) devices configured to measure the intensity of electric fields conducted by finger ridges and valleys to sense and capture the fingerprint image. Regardless of the method used to sense the fingerprint, conventional devices and techniques have common drawbacks, particularly when used in combination with portable electronic devices. These devices require small component size because of a lack of space and surface area due to the devices small size, and further require that any power demand be as small as possible due to limited battery life.
Specifically, devices exist that have a sensing area that is smaller than the fingerprint area to be imaged. Such devices are greatly desired because they take up much less space than a full fingerprint sensor. This is a very useful feature for small portable devices. These sensing devices generally consist of one or more imaging lines disposed perpendicular to the axis of motion. As the finger surface is moved across the sensor, portions of the fingerprint are sensed and captured by the device. These portions are subsequently reconstructed in a mosaic or overlapping manner. In operation however, current conventional devices have severe drawbacks. They generally require extensive processing resources for computing the algorithms and required data for reconstructing fingerprints.
For applications of fingerprint identification devices in portable electronics, such as laptops and cellular telephones, low power consumption is a strict requirement. Therefore, it is important to maintain minimal computation processing in such applications. Again, present conventional fingerprint sensor technology requires a substantial amount of processing, and thus requires a large amount of power to perform the required tasks for reconstructing fingerprints for identification. One major problem is that a large amount of pixel information is required to be recorded and matched in a short a mount of time, burdening the device processor and consuming substantial power. This is a big problem with small devices, which already have restrictions on power consumption.
One conventional device is described in U.S. Pat. No. 6,002,815 of Immega, et al. The technique used by the Immega device is based on the amount of time required for the finger to travel a fixed distance between two parallel image lines that are oriented perpendicular to the axis of motion. After a time history of samples are captured, the speed is determined by finding the time delay that provides the best match between data from the first line and data to from the second line. The device captures the entire image of an object and stores the image line by line. Such an object is illustrated as a photo copy of a document, and the reference does not suggest a fingerprint or other image. Thus, it is directed to a device and method for scanning an image passing over a perpendicular slit pair at a variable speed, as opposed to objects that pass over the slit pair at a fixed speed. It does not address the problem of excessive processor power expended to perform the process. Also, the perpendicular lines of the image are used for determining the speed of the object as it passes through the perpendicular slit where the image is captured. These recorded lines are also used in reconstructing the image when the scan is complete. Thus, a large amount of data is processed and stored in the process. The amount of processing resources required to calculate the speed at any given moment is immense, where the resources include time required, calculation by the processor and power demanded by the processor. Furthermore, this time series approach has the disadvantage that it is not possible to quickly determine an absolute distance of motion by comparing only the instantaneous data from the two image lines. This is true for all cases other than for the rare coincidental case where the finger happens to travel exactly the distance between the image lines during the interval between the two samples. Another problem arises when the object is moving much slower than the sample rate of the device. In this case, the number of samples needed to find a match is substantial. In addition, at slow speeds, the device must compare a larger number of stored lines in order to find a match. This greatly increases the computational requirements, placing a substantial burden on the device processor. Thus, expensive high order processors are required for adequate performance and substantial power is needed to operate such processors.
Another technique is described in U.S. Pat. No. 6,289,114 of Mainguet. A device utilizing this method reconstructs fingerprints based on sensing and recording images taken of rectangular slices of the fingerprint and piecing them together using an overlapping mosaic algorithm. Like Immega, the technique described in Mainguet is also computationally burdensome on the device processor. Furthermore, the Mainguet method requires a substantial amount of memory as well as a larger number of imaging pixels in order to properly record the images. Again, this method demands substantial power to perform algorithms, a big problem for power rationed portable devices.
For accurate fingerprint capture, it is often advantageous to provide a navigation function with the same device used for fingerprint sensing. The navigation function can provide more functionality in as little area as possible in a portable device, and provide a more accurate fingerprint image. However, conventional devices and methods for navigation require substantial processor resources, and thus demand more power. In such devices, in order to sense finger motion, the sensing device must sample the image at a periodic rate that is fast enough to ensure that a moving feature will be sampled when it passes both the primary imaging line and the auxiliary line of pixels. As a consequence, the sensor needs to operate at full imaging speeds, thus consuming full imaging power while in the navigation mode. Consequently, conventional navigation methods demand substantial power, and are thus impractical for small devices.
Thus, there exists a great need in the art for a more efficient means to accurately sense and capture fingerprints on portable devices and also to provide navigation operations without unduly demanding power. As will be seen, the invention provides a means to overcome the shortcomings of conventional systems in an elegant manner.